| 1 |
The following private members can be used by members of parser classes |
| 2 |
generated by b(). All data members are actually protected |
| 3 |
members inherited from the parser's base class. |
| 4 |
itemization( |
| 5 |
it() bf(size_t d_acceptedTokens_):nl() |
| 6 |
Counts the number of accepted tokens since the start of the tt(parse()) |
| 7 |
function or since the last detected syntactic error. It is initialized |
| 8 |
to tt(d_requiredTokens_) to allow an early error to be detected as |
| 9 |
well. |
| 10 |
|
| 11 |
it() bf(bool d_actionCases_):nl() |
| 12 |
When the bf(debug) option has been specified, this variable (bf(false) |
| 13 |
by default) determines whether the numbers of the case-entries that |
| 14 |
are about to be executed by the parser's tt(executeAction) member are |
| 15 |
displayed to the standard output stream. |
| 16 |
|
| 17 |
it() bf(bool d_debug_):nl() |
| 18 |
When the bf(debug) option has been specified, this variable (bf(true) |
| 19 |
by default) determines whether debug information is actually |
| 20 |
displayed. |
| 21 |
|
| 22 |
it() bf(LTYPE_ d_loc_):nl() |
| 23 |
The location type value associated with a terminal token. It can be |
| 24 |
used by, e.g., lexical scanners to pass location information of a |
| 25 |
matched token to the parser in parallel with a returned token. It is |
| 26 |
available only when bf(%lsp-needed, %ltype) or bf(%locationstruct) has |
| 27 |
been defined. nl() |
| 28 |
Lexical scanners may be offered the facility to assign a value to this |
| 29 |
variable in parallel with a returned token. In order to allow a |
| 30 |
scanner access to bf(d_loc_), bf(d_loc_)'s address should be passed |
| 31 |
to the scanner. This can be realized, for example, by defining a |
| 32 |
member bf(void setLoc(LTYPE_ *loc)) in the lexical scanner, which is |
| 33 |
then called from the parser's constructor as follows: |
| 34 |
verb( |
| 35 |
d_scanner.setSLoc(&d_loc_); |
| 36 |
) |
| 37 |
Subsequently, the lexical scanner may assign a value to the parser's |
| 38 |
bf(d_loc_) variable through the pointer to bf(d_loc_) stored inside |
| 39 |
the lexical scanner. |
| 40 |
|
| 41 |
it() bf(size_t d_nErrors_):nl() |
| 42 |
The number of errors counted by tt(parse()). It is initialized by the |
| 43 |
parser's base class initializer, and is updated while tt(parse()) |
| 44 |
executes. When tt(parse()) has returned it contains the total number |
| 45 |
of errors counted by tt(parse()). Errors are not counted if suppressed |
| 46 |
(i.e., if tt(d_acceptedTokens_) is less than tt(d_requiredTokens_)). |
| 47 |
|
| 48 |
it() bf(size_t d_requiredTokens_):nl() |
| 49 |
Defines the minimum number of accepted tokens that the tt(parse()) |
| 50 |
function must have processed before a syntactic error can be |
| 51 |
generated. |
| 52 |
|
| 53 |
it() label(DVAL) bf(STYPE_ d_val_):nl() |
| 54 |
The semantic value of a returned token or nonterminal symbol. With |
| 55 |
nonterminal tokens it is assigned a value through the action rule's |
| 56 |
symbol bf($$). Lexical scanners may be offered the facility to assign |
| 57 |
a semantic value to this variable in parallel with a returned |
| 58 |
token. To allow a scanner access to bf(d_val_), |
| 59 |
bf(d_val_)'s address should be passed to the scanner. This can be |
| 60 |
realized, for example, by defining a member bf(void setSval(STYPE_ |
| 61 |
*)) in the lexical scanner, which is then called from the parser's |
| 62 |
constructor as follows: |
| 63 |
verb( |
| 64 |
d_scanner.setSval(&d_val_); |
| 65 |
) |
| 66 |
Subsequently, the lexical scanner may assign a value to the parser's |
| 67 |
bf(d_val_) variable through the pointer to bf(d_val_) stored inside |
| 68 |
the lexical scanner. |
| 69 |
|
| 70 |
Note that in some cases this approach em(must) be used to make |
| 71 |
available the correct semantic value to the parser. In particular, |
| 72 |
when a grammar state defines multiple reductions, depending on the |
| 73 |
next token, the reduction's action only takes place following the |
| 74 |
retrieval of the next token, thus losing the initially matched token |
| 75 |
text. As an example, consider the following little grammar: |
| 76 |
verb( |
| 77 |
expr: |
| 78 |
name |
| 79 |
| |
| 80 |
ident '(' ')' |
| 81 |
| |
| 82 |
NR |
| 83 |
; |
| 84 |
|
| 85 |
name: |
| 86 |
IDENT |
| 87 |
; |
| 88 |
|
| 89 |
ident: IDENT ; |
| 90 |
) |
| 91 |
Having recognized tt(IDENT) two reductions are possible: to tt(name) |
| 92 |
and to tt(ident). The reduction to tt(ident) is appropriate when the |
| 93 |
next token is tt(CHAR(40)), otherwise the reduction to tt(name) is |
| 94 |
performed. So, the parser asks for the next token, thereby |
| 95 |
destroying the text matching tt(IDENT) before tt(ident) or tt(name)'s |
| 96 |
actions are able to save the text themselves. To enure the |
| 97 |
availability of the text matching tt(IDENT) is situations like these |
| 98 |
the em(scanner) must assign the proper semantic value when it |
| 99 |
recognizes a token. Consequently the parser's tt(d_val_) data member |
| 100 |
must be made available to the scanner. |
| 101 |
|
| 102 |
If tt(STYPE_) is a wrapper type for polymorphic semantic values, then |
| 103 |
direct assignment of values to tt(d_val_) is is only possible from |
| 104 |
values of the defined polymorphic data types. More complex assignments |
| 105 |
are also possible, using em(tagged assignments)+IFDEF(MANUAL)(, as |
| 106 |
explained in section ref(POLYTYPE))(). |
| 107 |
) |